The Cost Problem Closed and a Pit Wall Opened: Agnico's Best Operating Quarter in a Year Arrives Into a Broken Gold Tape

Results vs. Consensus

Q2 2026 Scorecard

Year-Over-Year Comparisons

Operating Metrics, Year Over Year

The 2025 cost figures above are the restated ones Agnico printed alongside the Q2 2026 numbers. Effective January 1, 2026 the company changed how total cash costs and AISC treat the NTI Payment at Meadowbank, and it has restated the 2025 comparatives onto that basis. On the old composition Q2 2025 total cash costs were $933 per ounce and AISC was $1,289. Anyone comparing this quarter against a figure lifted from last year's release is looking at two different measures.

Quarter-Over-Quarter Comparisons

Both quarters sit on the same cost composition, so the sequential comparison needs no adjustment. It is the cleanest read available on whether the first quarter was noise or trend, and the answer is noise: production up 3.7%, both unit-cost measures down, and the free cash flow line up 82.4% as the $1.3 billion 2025 tax catch-up washed out of the base.

Quality of Beat

• Revenue. A 35.0% increase on 1.3% fewer ounces sold. The realized price did all of it and then some: $4,483 against $3,288 a year ago is $1,195 per ounce of pure price, worth roughly $998M on 835,505 ounces sold, against a reported revenue increase of $986.7M. Volume was a small negative and by-products a small positive. There is no organic revenue growth in this print and management does not claim any.
• Margins. Gross margin expanded 520bp to 63.8%, but the interesting margin fact is compositional. Total cash costs of $1,054 include a $75 per ounce by-product credit against $46 a year ago. Strip the incremental $29 and the quarter prints $1,083, which is above the $1,070 guidance midpoint rather than below it. The CFO named this himself, crediting a stronger US dollar, lower royalties on a lower gold price, and conservative copper and silver assumptions. Roughly half of the cushion below the midpoint is commodity and currency arithmetic rather than productivity, which does not make it less real but does make it less repeatable.
• EPS. IFRS net income of $1,600.5M includes a $155.3M pre-tax gain on the sale of investments, worth $0.31 per share after tax, and $56M of net derivative losses, worth $0.11. Adjusted net income of $1,541M strips both, and the $3.07 basic adjusted figure is the clean one. The tax line is unremarkable at an effective rate consistent with the 34% to 36% full-year guide. The below-the-line noise is larger than usual this quarter and it runs in both directions.
• Free cash flow. The $1,335M headline is a company record and is also the least informative number in the release. Operating cash flow rose 16.2% and capital spending rose 48.9%, so the record is a working-capital artefact: the year-ago quarter drew $513M out of working capital and this one drew $32M. Before working capital, free cash flow went from $792M to $1,303M, up 64.5%. That is the number that describes the business.

Segment Performance

Agnico does not publish revenue by region in the quarterly release, so the table below uses payable production and the regional total cash cost per ounce, both of which the company does disclose per mine and per region. The 2025 cost figures are on the restated composition, so the comparison is like for like.

Production and Unit Costs by Region

Canada supplied 85.4% of payable production, which is the jurisdictional pillar of this thesis expressed in ounces rather than dollars. The regional mix shifted meaningfully within that: Ontario overtook Quebec as the largest producing region, a direct consequence of Detour Lake growing 23.2% while Canadian Malartic shrank 21.6%.

Production and Unit Costs by Mine

Three of ten mines grew production year over year and four of ten produced gold more cheaply. In the first quarter the comparable counts were three of ten and one of ten. The distribution has improved, and it has improved in the places that matter: Detour Lake is 24.2% of consolidated production and took its unit cost down 9.7%, while Kittila is the only mine in the portfolio to have held its cost flat across a year of double-digit sector inflation.

Abitibi Ontario — the quarter's engine

Detour Lake produced 207,279 ounces at $825 per ounce, the lowest unit cost in the portfolio, on a record 30.6 million tonnes of ore and waste extracted and a record 80,275 tonnes per day through the mill at approximately 96% runtime. The grade sequence helped, and management was explicit that it will not hold: the profile was scheduled higher in the first half. Macassa delivered a second consecutive quarterly mill throughput record at 2,593 tonnes per day, and began trucking ore from the AK deposit to the LZ5 facility at LaRonde, processing 71,000 tonnes for 7,800 ounces.

"We achieved another consecutive quarterly record in tonnes mined, and this is the result of the productivity initiatives started last year, which are now paying off. An example is increasing shovel utilization by increasing the amount of blasted material inventory that's available, and also getting that higher shovel productivity by diligently improving on our loading practices." — Natasha Vaz, EVP and COO, Ontario, Australia and Mexico

Macassa's unit cost remains the portfolio's worst year-over-year comparison at +66.3%, though the base is flattering: $626 per ounce in Q2 2025 was an exceptional print. At $1,041 the mine now runs marginally below the consolidated average, and the first-half figure of $1,129 is the more representative number. The AK contribution of 7,800 ounces represents 19.5% of the roughly 40,000 ounces the deposit was expected to add in 2026, achieved in the first quarter of trucking.

Assessment: Ontario carried the quarter and it did so on process improvement rather than price, which is the only kind of improvement that survives a gold retracement. The caveat is grade: Detour's first-half sequence was the high one, and management confirmed the grade comes off in the second half while throughput holds. Expect Ontario's cost advantage to compress in Q3 and Q4 without anything going wrong.

Abitibi Quebec — where the damage is

Quebec produced 245,781 ounces, down 17.2%, at $1,096 per ounce, up 26.9%. Canadian Malartic is the entire story: production fell 21.6% to 135,243 ounces on lower throughput and grade, compounded by an unscheduled six-day mill shutdown related to the fatal accident in April. Odyssey, the underground operation being built inside the same complex, delivered a record 28,800 ounces, or 21.3% of Malartic's total. LaRonde ran ore tonnes above plan on positive reconciliation in the East Mine, and Goldex was in line.

"In Quebec, Canadian Malartic are facing more challenges, but thanks to the team led by Dan, Serge, and Justin for their management of those challenges and their dedication to it. Overall, Canadian Malartic finished the first half on target, unless they have some challenges." — Dominique Girard, EVP and COO, Nunavut, Quebec and Europe

The July 1 rock mass movement along Barnat's north wall arrived after the quarter closed but reframes everything about this region. Canadian Malartic's full-year total cash cost guidance moved to approximately $1,260 per ounce from $1,187, a 6.1% step-up, and the mine will not be mining Barnat at all in the third quarter.

Assessment: Quebec is now a two-speed region. Odyssey is ramping ahead of plan and is the asset the 2030s thesis is built on. Barnat is an end-of-life open pit that just lost roughly 370,000 ounces and will spend a quarter building berms. The company's own framing is that the complex is transitioning from pit to underground, and this event accelerates a transition that was always coming. It also removes the low-cost tonnage that was subsidising the transition.

Nunavut — flat production, rising costs, sealift risk ahead

Nunavut produced 197,681 ounces, up 2.9%, at $1,121 per ounce, up 8.9%. Meliadine set a quarterly mill throughput record of 648,000 tonnes, averaging 7,121 tonnes per day against a 6,500 tonne target, with gold production in line as higher tonnes offset lower grades. Meadowbank came in below plan on mine sequencing and grade, with freshet conditions and heavier-than-expected June rainfall constraining open-pit ore mined. Full-year Meadowbank production is described as in line with plan.

"For Nunavut, the first half of the year production is also on plan. Good news: the spring migration is over, and it went very well. To date, 6 of the 19 vessels are already received from our sealift." — Dominique Girard, EVP and COO, Nunavut, Quebec and Europe

Assessment: The region's cost inflation is the most structural in the portfolio because Nunavut is the diesel-exposed part of an otherwise hydro-and-nuclear-powered asset base. The company has now bought roughly 70% of its Nunavut diesel requirement for the balance of 2026 and through the 2027 sealift, about 130 million litres, at prices roughly 30% above budget net of hedges. That purchase is a known, sized, already-incurred cost increase, and it is the single most quantifiable cost headwind in the release.

Finland — the best-run asset in the portfolio this quarter

Kittila produced 61,969 ounces, up 23.1%, at $1,133 per ounce, down 0.1%. Record mill throughput of 6,407 tonnes per day and positive grade reconciliation drove the volume; the flat unit cost across a year in which the consolidated figure rose 13.9% is the more remarkable number. Management stated that first-half 2026 unit costs, excluding royalty and mining tax, were lower than first-half 2025, which were themselves lower than 2024.

"Maintaining a declining unit cost trend despite inflationary pressures and a deepening underground operation is not easy and is clear evidence that systematic productivity work makes a difference." — Jani Lösönen, VP, Europe

Mill recovery remained at 81%, below plan, with four named improvement initiatives running and no target or date attached to any of them. This is the second consecutive quarter in which the recovery shortfall was disclosed and the second consecutive quarter in which no analyst asked about it.

Assessment: Kittila is the proof case for the productivity argument management makes across the whole portfolio, and it is now also the operating anchor for a Finnish platform meant to reach 500,000 ounces a year. The 81% recovery is the one unresolved operational item, and it is worth roughly the difference between a good mine and an excellent one at a $4,000 gold price.

Australia and Mexico — small, and moving in opposite directions

Fosterville produced 42,012 ounces, down 15.3%, at $1,114 per ounce, up 42.3%. Production beat plan on higher grades from positive reconciliation and better recovery, offset by lower throughput while the underground primary ventilation upgrade was commissioned. Development rates are running 14% ahead year to date. Pinos Altos produced 20,951 ounces at $1,873, the portfolio's highest unit cost but its second-best year-over-year cost change at -6.4%, helped by a large by-product credit.

Assessment: Fosterville's cost line looks alarming and is mostly denominator: a 15.3% production decline against a largely fixed cost base while ventilation capital is spent. The mill upgrade targets 3,300 tonnes per day from 2028, so the ounces come back. Neither asset is large enough to move the consolidated thesis, and together they are 7.4% of production.

Key Topics & Management Commentary

Overall Management Tone: Confident and unusually specific on operations, and noticeably more comfortable than at the Q1 call, where the "tracking to plan" framing had to carry a quarter whose cost lines contradicted it. This time the numbers did the work and the commentary could be about execution detail: automated haulage productivity, mill runtime, development metres. The one place management declined to engage was the 2027 cost effect of the Barnat event, deferred to a budgeting process not yet complete, and any framework for what the Rupert contingent value rights are worth. The safety section was the most substantive we have heard from this company, and it was delivered before the financial results rather than after.

1. The Barnat Wall: 370,000 Ounces, Three Years, and a Berm

On July 1, approximately one million tonnes of rock moved along the final north wall of the Barnat open pit at Canadian Malartic. The area was already under enhanced geotechnical monitoring and mining had been suspended pre-emptively, so there were no injuries, no equipment damage and no environmental impact. The cause is attributed to weaker altered rock and structures associated with the Cadillac Fault, with ground conditions worsened by freshet and heavy rainfall. The moved material stays where it is. A safety catchment area and berms 15 to 25 metres high go in during the third quarter, a temporary access ramp along the southwest wall is already built, and mining resumes in the fourth quarter.

"For example, on July 1st, we had a rock movement in the wall of our Barnat pit. Of course, this was a disappointment. But I am proud of our team and, importantly, of our systems and our processes, including the systems and processes we had in place to track potential wall movement, that allowed us to move quickly to protect both our people and our equipment." — Ammar Al-Joundi, President and CEO

The production effect is 60,000 to 80,000 ounces in the second half of 2026 and up to 150,000 ounces in each of 2027 and 2028. On the call the operating executive put the total sterilised at roughly 370,000 ounces against roughly 300,000 ounces still mineable at Barnat, and Canadian Malartic's full-year unit cost guidance moved to approximately $1,260 per ounce from $1,187. Low-grade stockpile feed partially fills the mill.

Assessment: Sized against a three-year production base near ten million ounces, 370,000 ounces is roughly 3.7%, which is material but not thesis-breaking. Two things make it worse than the headline. The 2027 and 2028 cost effect has not been quantified and will not be until the budget cycle completes. And the ounces removed are cheap open-pit ounces at the company's second-largest Quebec asset, so the mix effect on consolidated unit costs runs against the improvement this quarter demonstrated.

2. Costs Came in Below Both Guidance Midpoints for the First Time This Year

Total cash costs of $1,054 per ounce and AISC of $1,459 both landed below their respective full-year guidance midpoints of $1,070 and $1,475. In Q1 both finished above. That reversal is the specific item this note flagged three months ago as the thing to watch, and it resolved in the company's favour by more than the margin required.

"Total cash costs were $1,054 per ounce, and all-in sustaining costs were $1,459 per ounce, below our costs in the first quarter, below the midpoint of our guidance ranges, and hundreds of dollars below the industry average." — Jamie Porter, EVP, Finance and CFO

The first-half figures are $1,073 and $1,471, both marginally above their midpoints. Holding the $1,070 full-year midpoint from here requires roughly $1,067 per ounce across the second half, which is $13 above what Q2 delivered. Breaching the $1,120 top of the range would take $1,167.

Assessment: The arithmetic is now comfortable rather than tight, which is a genuine change from May. The complication is that the Barnat remediation quarter and the diesel purchased 30% above budget both land in the second half, and Canadian Malartic's own guide implies its costs rise from $1,185 in Q2 toward $1,260 for the year. The cushion exists; it is not large.

3. About Half the Cost Cushion Is Currency and By-Products

The by-product credit inside total cash costs was $75 per ounce, against $46 a year ago. Add back the incremental $29 and the quarter prints $1,083, above the guidance midpoint rather than below it. The CFO volunteered the decomposition rather than being pressed on it.

"But we are seeing a much stronger US dollar than what we budgeted and guided at the start of the year. So we're getting the benefit from that. You'll recall that we also guided at a $4,500 gold price. So we're seeing a bit of a benefit on cash costs with respect to royalties. And thirdly, with respect to by-product credits, we have more conservative assumptions on our copper and silver pricing. So those are all helping." — Jamie Porter, EVP, Finance and CFO

There is a symmetry here worth noting. Agnico's cost guidance assumes $4,500 gold, and royalties and the NTI Payment scale with the realized price. A falling gold price therefore lowers the cost line at the same time it lowers revenue. That mechanism ran in reverse through 2025, when costs finished above the guided range while production landed above the midpoint.

Assessment: This is the honest reading of the beat: perhaps half the cushion below the midpoint is productivity and the other half is a strong dollar, a weak gold price and conservative by-product assumptions. Management said so unprompted, which is a mark in its favour. It also means the cost improvement is partly a hedge against the thing hurting the equity, and partly not repeatable if the dollar turns.

4. Record Free Cash Flow, and What "Record" Is Actually Measuring

Free cash flow of $1,335M is the highest quarterly figure in the company's history and is 2.3% above the year-ago quarter. Those two facts sit awkwardly together, and the reconciliation is working capital: the June 2025 quarter released $513M from working capital, this one released $32M. Before working capital, free cash flow went from $792M to $1,303M, up 64.5%.

"Strong operational performance and disciplined cost management combined with a favourable gold price environment to drive record free cash flow of over $1.3 billion for the quarter." — Jamie Porter, EVP, Finance and CFO

Operating cash flow rose 16.2% to $2,144M. Capital spending including capitalised exploration rose 48.9% to $801M, of which $143M went to Hope Bay. Cash taxes of $623M were paid in the quarter, bringing the first-half total to $2.4 billion, roughly 70% of the full-year expectation.

Assessment: The record label is technically correct and analytically empty. What matters is that the tax profile is now front-loaded and done: roughly $1.1 billion remains for the second half against $2.4 billion paid in the first, which is a mechanical tailwind of about $600M per quarter to reported free cash flow regardless of what gold does.

5. The Capital Program Grew by Roughly $415M

Total 2026 capital expenditures excluding capitalised exploration are now guided to $2.6 to $2.8 billion, against $2.2 to $2.4 billion previously. The increase is entirely the Hope Bay construction sanction announced May 19. Capitalised exploration guidance is unchanged at $290 to $330 million, so total capital including exploration now runs $2.9 to $3.2 billion against $2.5 to $2.7 billion before.

"This quarter, we invested over $800 million in advancing key projects and in capitalized exploration." — Ammar Al-Joundi, President and CEO

Development capital ran 1.95 times sustaining capital in the quarter, against 1.49 times in Q1. The revolver remained undrawn and available liquidity stayed at approximately $2 billion, excluding an uncommitted $1 billion accordion.

Assessment: A $415M capital increase against a realized gold price falling $378 per ounce sequentially is the pinch point in this story. Both moves are individually defensible and together they compress free cash flow from two directions at once. The company can afford it: net cash rose anyway. But the "funded from operating cash flow" claim in the growth thesis has less headroom at $4,036 gold than it did at $4,861.

6. Hope Bay Sanctioned at Roughly $2.4 Billion, Near the Top of the May Framing

The construction decision arrived May 19, on schedule. The study envisions 400,000 to 435,000 ounces a year over an initial 11-year mine life, with average total cash costs of approximately $958 per ounce and AISC of approximately $1,214, on a $4,500 gold price and a 1.36 Canadian dollar. Initial development capital is approximately $2.4 billion. Detailed engineering reached approximately 67% at quarter end, and management put it above 70% on the call, three new camp wings are complete, and nine sealift vessels are scheduled, with the first leaving Becancour in early August.

"We've announced the go-ahead of our Hope Bay mine. This will be a world-class, low-cost mine producing between 400,000 and 450,000 ounces a year that we expect to happen for decades." — Ammar Al-Joundi, President and CEO

Three months ago the range in circulation was "slightly over $2 billion" against "below $2.5 billion." The number landed at the top of that band.

Assessment: Hope Bay's projected AISC of roughly $1,214 per ounce would sit below the current consolidated figure, so the project is accretive to unit costs when it arrives. It is also the largest single capital commitment in the company's pipeline, sanctioned into a gold tape that has fallen sharply from its spring high, and the economics are quoted at a $4,500 gold assumption that no longer holds. The project does not need $4,500 gold to work at a $958 cash cost. It does need the balance sheet, which is why the net cash pillar matters more this quarter than last.

7. The Buyback Quadrupled, and It Bought at $178.86

Agnico repurchased 2,235,947 shares in the quarter at an average of $178.86, for $400 million, against $150 million in Q1. Combined with the $0.45 quarterly dividend, total shareholder returns were $625 million, a company record, and 48% of first-half free cash flow against a 40% target. Management flagged the potential to exceed the full-year target. The repurchases were part-funded by $261 million of equity portfolio sales, which is precisely the mechanism the company named in April as its way of offsetting the Rupert dilution.

"Beginning of the year, we set a target of returning approximately 40 percent of free cash flow to shareholders. Through the first half of the year, we've exceeded that objective, returning approximately 48 percent of free cash flow through dividends and share repurchases." — Jamie Porter, EVP, Finance and CFO

The average purchase price of $178.86 compares with a June 30 close of $155.13 and a July 29 close of $144.43. The buyback was executed 15.3% above where the stock finished the quarter.

Assessment: The commitment was made in April, dated, and delivered on the number, which is what a scorecard needs and is genuinely to management's credit. The execution is another matter. Buying $400 million of stock at an average 15% above the quarter-end price is value-destructive on a mark-to-market basis, and the policy that produced it, a fixed percentage of free cash flow, is structurally price-insensitive. A company that talks this much about per-share value creation should be buying more when the stock is cheaper, not the same amount regardless.

8. Safety: A Third Fatality, and the First Time It Cost Ounces

Daniel Giroux died at Upper Beaver on May 1. It is the third fatality in twelve months, after Fosterville in December 2025 and Canadian Malartic in April 2026. The CEO opened the call's substance with it, before the financial results, and gave the count directly: 23 fatalities across almost seventy years of operation, three of them in the last year.

"In the almost-70 years of operation, from 1957 to today, we've had a total of 23 fatalities, and three of these have occurred in the last year. I will repeat what I said last quarter. Fatalities, every single one, is not acceptable." — Ammar Al-Joundi, President and CEO

A global safety reset was implemented across all operations. The company is accelerating identification and implementation of critical controls, strengthening supervision, and rolling out a supervision training programme. The concrete financial consequence appeared this quarter for the first time: the six-day unscheduled Canadian Malartic mill shutdown related to the April accident is named in the release as a driver of that mine's production shortfall. The ounce cost is not quantified.

Assessment: Three fatalities in a year at a company whose cost advantage is explicitly attributed to stable, long-tenured workforces is now a thesis risk rather than a governance footnote, and it has started to show up in production. The disclosure quality is high and the root-cause detail given on the call was specific and unflattering to the company in places, which is the correct posture. That does not change the trend line.

9. Detour and Kittila Delivered, and the Records Claim Checks Out

The CEO opened with a specific, verifiable claim: record mill throughput at mines representing slightly more than half of total production. Detour Lake at 80,275 tonnes per day, Meliadine at 7,121, Kittila at 6,407 and Macassa at 2,593 together produced 446,907 ounces, which is 52.2% of the consolidated total. The claim is accurate.

"Individually, these may seem like small accomplishments, but when we step back and when we look at the big picture collectively, this quarter, we've had record mill throughput at mines representing slightly more than half of our total production." — Ammar Al-Joundi, President and CEO

Underneath it, the automation detail was the most concrete operating disclosure of the call. At LZ5, fully automated shifts run Friday, Saturday and Sunday nights, productivity is up 65% in the first half at roughly 2,030 tonnes per fully automated shift, and interventions requiring the sequence to stop fell from 1,700 per shift to 700. At LaRonde, roughly 25% of ore mucking in the first half used automated loaders.

Assessment: This is the evidence base for the claim that productivity can offset inflation, and it is unusually specific by the standards of gold-sector conference calls. The CFO put a number on the historical rate: roughly 7% average inflation over three years against 3% to 4% cost growth excluding royalties, so about half offset. That is a defensible track record and it is the reason to treat the Q2 cost print as partly structural rather than entirely a currency accident.

10. Finland Closed, and the Share Count Steps Up in the Third Quarter

All three Finnish transactions completed in the quarter: B2Gold's 70% of the FinGold joint venture for $325 million cash on April 22, Aurion for $339 million cash on June 15, and Rupert on June 16 for 0.0401 Agnico shares each, aggregate consideration of $1,687 million, plus contingent value rights. The consolidated land package is 2,492 square kilometres and 44 employees transferred. Cash out the door on the Finnish acquisitions was $578 million in the quarter.

"In Q2, the time was right for a more significant consolidation transaction, which will solidify Agnico's position in Finland for decades to come." — Jani Lösönen, VP, Europe

Common shares issued rose from 500,653,224 at March 31 to 506,899,374 at June 30, a net increase of 6.25 million after retiring 2.24 million to the buyback, implying gross issuance of roughly 8.5 million shares. The Q2 weighted average basic share count was 501.7 million because the Rupert shares were issued on June 16. Period-end shares net of those held in trust are 506.4 million, so the run-rate count is roughly 0.9% above the figure that divided this quarter's earnings.

Assessment: The dilution is now measurable and it is modest, about 1.2% on issued shares net of the buyback against a land package the company argues is a path to a 500,000 ounce regional platform. The first optimisation study is due at the end of 2027, so there is no interim mark. What has not been disclosed is the carrying value of the 207,654,166 contingent value rights, each worth up to $3.00 on reserve and production milestones over ten years. The maximum aggregate exposure is roughly $623 million and the release does not state what has been recognised.

11. Diesel Bought 30% Above Budget for the Nunavut Balance of Year

With the 2026 sealift underway, Agnico has purchased approximately 70% of its Nunavut diesel requirement for the balance of 2026 and through the 2027 sealift, roughly 130 million litres, at prices approximately 30% above the budgeted $0.78 per litre benchmark, net of hedges. Diesel is approximately 10% of operating costs, 7% direct and 3% transportation. The company estimates a 10% change in diesel prices moves total cash costs by approximately $4 per ounce in the second half, and a further $2 through transportation.

"And we do have some of that exposure hedged for the back half of this year, but that's going to be, I'd say, the biggest kind of cost pressure in 2027 relative to 2026. And diesel, just as a reminder, represents about 7 percent of our overall cost." — Jamie Porter, EVP, Finance and CFO

The CEO framed the quarter's cost performance against the same input: gold production above budget with costs inside the guidance range "in a quarter where oil traded above $100 per barrel for much of the time."

Assessment: This is a known, sized, already-incurred headwind, which makes it the most tractable line item in the release. Applying the company's own sensitivity, a 30% overrun is worth roughly $12 to $18 per ounce on second-half total cash costs, against a $13 cushion. The diesel purchase alone consumes most of the room between Q2's print and the guidance midpoint.

12. 2027 Is a Blank Page

Barnat removes up to 150,000 ounces in each of 2027 and 2028 and pushes Canadian Malartic's unit costs higher through both years. Asked directly what that does to the portfolio, management confirmed the direction and declined the magnitude, on the grounds that the multi-year budgeting and planning process is only now underway. The only forward number offered was an inflation assumption: 3% to 4% for labour and contractors, which are 40% to 50% of the cost structure.

"So it's premature to be commenting on future-year costs. But we obviously will see slightly higher costs at Canadian Malartic, given less production through '27 and '28. But on an overall basis, again, we'll do what we can to offset inflation through continuous improvement and other efficiency initiatives." — Jamie Porter, EVP, Finance and CFO

The offsets named were qualitative: continuous improvement, efficiency initiatives, and a track record of absorbing roughly half of input inflation.

Assessment: Reasonable in July and unacceptable by February. The event happened on July 1, the production effect was quantified within 24 hours, and the cost effect is a function of the same mine plan. The February reserve and guidance release is the deadline, and it is also where the 370,000 sterilised ounces have to be dealt with in the reserve statement. Both are now scheduled disclosures rather than open questions.

Guidance & Outlook

Only two things changed, and they point in opposite directions with respect to the equity. Production skewed toward the lower end of an unchanged range because of Barnat, and capital rose $415 million at the midpoint because Hope Bay was sanctioned. Everything else, including both unit-cost ranges, is exactly where it was in February. The guidance assumptions underneath are also unchanged: $4,500 per ounce gold, $0.78 per litre diesel, 1.36 Canadian dollars, 1.18 US dollars per euro, 1.40 Australian dollars and 17.50 pesos. Spot gold on July 29 closed at $4,036.30, 10.3% below the assumption.

Implied second-half shape. Against first-half production of 1,680,925 ounces, a full-year outcome of 3,350 thousand ounces implies roughly 1,669 thousand ounces in the second half, an average of 835 thousand per quarter. That is essentially flat with Q2 and requires the Barnat gap to be filled by low-grade stockpile throughput at Malartic plus continued outperformance at Detour, Kittila and Fosterville, in a half that carries scheduled shutdowns at LaRonde, Detour Lake, Macassa, Meliadine, Meadowbank and a 17-day autoclave reline at Kittila.

Implied second-half cost arithmetic. Holding the $1,070 total cash cost midpoint requires approximately $1,067 per ounce across the second half, $13 above Q2's print. The $1,120 top of the range would not be reached until $1,167. On AISC the arithmetic is tighter: holding the $1,475 midpoint needs roughly $1,479, which is $20 above Q2, and sustaining capital per ounce has been rising, at $286 in Q2 against $265 for the half.

Street at: The compiled estimates going into the print sat at 827,779 ounces and $1,043 per ounce for the quarter, so the Street was modelling weaker volume and tighter costs than the company delivered. On the full year, the change that matters to estimates is not the production skew, which was disclosed July 2, but the capital increase, which was not in numbers before this release.

Guidance style: Agnico guides ranges and then narrows the skew rather than moving the range, which it has now done twice this year in opposite directions. It also holds cost guidance through events that clearly affect one mine's costs, absorbing them at the portfolio level rather than re-cutting the number. The pattern in 2025 was full-year costs finishing above the top of the guided range while production landed above the midpoint, which is the risk case here if the gold price recovers and drags royalties with it.

Analyst Q&A Highlights

How Much Gold the Barnat Wall Actually Took Out

The first question of the session went straight past the remediation plan to the resource consequence, asking for the volume and grade of material now inaccessible. The answer was more forthcoming than the release: a specific sterilised ounce count, a specific split across years, and a specific remaining figure for the pit, offered without hedging and with an explicit admission of what was not known.

Q: "Just going back to Barnat for a moment, you talked about the 1 million tonnes of material that slid. Is there any way the Company can quantify the volume and grade of material that would be inaccessible as a result of that slip?"
— Josh Wolfson, RBC Capital Markets

A: "Well, we disclose that there's 370,000 ounces that will not be accessible anymore, which is 60,000, 80,000 in '26, 150,000 in '27/'28. Let's say the remaining ounces in Barnat is approximately 300,000 ounces that we're going to mine."
— Dominique Girard, EVP and COO, Nunavut, Quebec and Europe

Assessment: The most useful ninety seconds of the call. It converts a production-guidance footnote into a resource number, and it establishes that more than half of what was left in Barnat is gone. At roughly 1.0 to 1.1 grams per tonne the sterilised ounces are low-grade open-pit material, which is why the volume effect is larger than the value effect. The follow-up, on whether low-grade stockpiles could offset, drew a straightforward yes on mill feed and a reminder that the pit was near end of life regardless.

Whether the Sterilised Ounces Ever Come Back from Underground

A separate line of questioning tested whether the 370,000 ounces could be recovered later through underground access once the open pit closes. The answer was a clear refusal to book anything, paired with a date on which a better view will exist. It is the kind of answer that is unhelpful in the moment and useful in a model.

Q: "Dominique, do you think that there's the potential at Canadian Malartic to come back got that 370,000 ounces that we've left behind, come back at it after, at the end of the mine life of the open pit, and access it from underground?"
— Tanya Jakusconek, Scotiabank

A: "Well, we are keeping understanding and redesigning the pit. But, Tanya, I will not put that in the book anywhere now that we're going to recover them. We might see opportunity with time, the years to come. I guess next February might have a better view, but I don't expect for now to recover those ounces."
— Dominique Girard, EVP and COO, Nunavut, Quebec and Europe

Assessment: Treat the 370,000 ounces as gone. The February reserve statement is the checkpoint, and it is also where the reserve accounting for those ounces has to be settled. Management is not planting an option here, which is the right posture and makes the number easier to model.

What the Malartic Cost Step-Up Does to the Rest of the Portfolio

The question that most directly tested the credibility of an unchanged cost guide: if one mine's full-year unit cost moves up 6%, what absorbs it? The answer named three specific offsets, all of them external to operations, and then declined the forward-year version of the same question.

Q: "If I could just circle back to Canadian Malartic, you took the cash cost guidance up for the rest of this year for the second half. Just wondering how we should think about cash costs across the portfolio to mitigate that. And then also, just going forward to '27 and '28, does the higher cost in the second half of this year have any impact going forward?"
— Richard Garchitorena, Barclays

A: "So it's premature to be commenting on future-year costs. But we obviously will see slightly higher costs at Canadian Malartic, given less production through '27 and '28. But on an overall basis, again, we'll do what we can to offset inflation through continuous improvement and other efficiency initiatives."
— Jamie Porter, EVP, Finance and CFO

Assessment: The three offsets named for 2026 (a stronger dollar, lower royalties on lower gold, conservative by-product assumptions) are all price and currency effects, not operating ones. That is a candid answer and it also concedes the point: the reason the consolidated cost guide survives a 6% increase at Malartic is largely that gold fell. If the gold price recovers, the royalty offset reverses at the same moment the revenue improves.

Whether Productivity Gains Can Actually Offset Inflation

A recurring line of questioning on the call asked management to put a number on the claim that continuous improvement absorbs input inflation. The response was the most quantified statement of the session and, unusually, it was a partial concession rather than a defence.

Q: "But we've talked a lot about this optimization and the productivity improvements that you're seeing, both mill and equipment. All else being equal, do you think that we can offset inflation with all of these optimizations? So for example, if inflation's 4 percent overall, do you think all of this can offset that, or partially, all else being equal?"
— Tanya Jakusconek, Scotiabank

A: "If you look back over the last three years, though, I'd say on average, inflation's probably run around 7 percent. And if you back out royalties going up because of higher gold prices, I'd say, on average, our costs have been up 3 percent to 4 percent. So over the last three years, we've offset almost half of the inflation through continuous improvement and productivity initiatives. So that, obviously, would be the target going forward."
— Jamie Porter, EVP, Finance and CFO

Assessment: A 50% offset rate over three years is a real, checkable track record and it is better than most of the sector. It is also an admission that half of input inflation does reach the cost line, which is precisely the mechanism behind the 13.9% year-over-year increase in total cash costs this quarter. The useful takeaway for the model is that unit costs should be expected to rise at roughly half the rate of input inflation, plus or minus the royalty effect from the gold price.

What the 2027 Inflation Assumption Looks Like

The only forward-year number offered on the call, extracted by asking about the budgeting process rather than about guidance. It arrived with the cost-structure weighting attached, which makes it usable.

Q: "Just curious if you could give us a look at what you're seeing as a reasonable inflation assumption going into 2027."
— Lawson Winder, Bank of America

A: "We're seeing CPI in Canada, that that's obviously going to impact our labour and contractor costs, which is 40 percent, 50 percent of our overall cost structure. So 3 percent, 4 percent for labour is probably not unreasonable at this time, but we'll see as we get closer to the end of the year."
— Jamie Porter, EVP, Finance and CFO

Assessment: Labour at 3% to 4% on 40% to 50% of the cost base is roughly $13 to $22 per ounce of gross 2027 pressure before offsets, against diesel described separately as the largest single 2027 headwind. Applying the company's own half-offset track record, a starting point of $1,090 to $1,110 per ounce for 2027 total cash costs is defensible before any Barnat mix effect. That mix effect is the piece nobody has sized.

Labour Availability in Northern Ontario

Prompted by a peer's commentary on rising labour and contractor costs in the same region, this exchange produced the operating answer first and then a strategic one that went considerably further than the question asked.

Q: "One of your peers talked about increased labour costs and contractor costs in Northern Ontario. And I saw in your release, there was a mention of just higher labour costs, but that's more year over year. Can you just touch on what the labour dynamics are looking like in Ontario specifically?"
— Fahad Tariq, Jefferies

A: "As you know, we have the lowest turnover of any of our peers. We probably have half the turnover of our peers. And actually, the turnover, over the last year, we've reduced it. But we're going beyond that. I'll give you some examples."
— Ammar Al-Joundi, President and CEO

Assessment: Internal labour cost inflation was put at roughly 4% year over year in Ontario, with no material contractor cost increase. The more interesting disclosure is the capital response: the company is building permanent homes in communities rather than camps, which is an unbudgeted-looking commitment that does not appear as a line item anywhere. Turnover at half the peer rate is the operational fact underneath the whole cost-advantage argument, and it is the thing three fatalities in a year could damage.

What the Three Fatalities Have in Common

The longest answer of the call, and the only one delivered by the sustainability executive. Asked for lessons learned, the response walked through all three incidents individually with root causes named. The follow-up below pressed on whether the remediation work could go faster, and drew a direct refusal to accelerate on demand.

Q: "And implementing all of this, Carol, like can it be done quickly?"
— Tanya Jakusconek, Scotiabank

A: "So a critical controls journey, we've been working with one of the experts in the world, quite frankly, on that for over a year now. And their advice to us is not to try to go too quickly. We can accelerate certain aspects of it."
— Carol Plummer, EVP, Sustainability, People & Culture

Assessment: The three root causes given were an unrecognised pinch point on a cable bolter that the equipment manufacturer had also not recognised, a decade-long erosion of engineering controls around a mill conveyor, and an experienced crew changing a work practice to mitigate a perceived risk without communicating the change for risk assessment. Those are three different failure modes, which is both reassuring (no single systemic defect) and not (three different failure modes in twelve months). The refusal to accelerate on demand is defensible and is also an answer that will not satisfy anyone underwriting the operational risk.

Whether the Early Shaft Completion Pulls Odyssey Forward

A detailed operational exchange on the Odyssey underground build, testing whether shaft sinking finishing three months early translates into an earlier commissioning date. It did not, and the answer explained why in a way that also surfaced a development-rate slip during the quarter.

Q: "Like the commentary says that the last bench was taken out on July 9th, and when I look back at the Q1 commentary, it was supposed to be, I guess, completed at the end of the year."
— Anita Soni, CIBC World Markets

A: "So we're ahead of schedule on the shaft sinking, but it doesn't mean that we're going to be faster and be better for Q2 next year for the commissioning. There's still lots of work to do. We might have a bit of a contingency with those advances, but it doesn't change the dates. We're well positioned. On the ramp development, we get a bit of delay this quarter, and the team is working to catch up on that in the coming quarters."
— Dominique Girard, EVP and COO, Nunavut, Quebec and Europe

Assessment: The correct answer, and the honest one. Shaft sinking is not the critical path; the loading station, headframe changeover and paste plant are. The disclosure worth extracting is the ramp development slip, currently running near 1,800 metres per month against a fourth-quarter target of 2,000, attributed in the release to higher ground support requirements and material handling at depth. Q2 2027 commissioning is unchanged and remains the single most important date in the Quebec growth plan.

How M&A Is Framed With the Sector De-Rated

With gold equities down sharply and the Finnish consolidation just closed, the question was whether the pullback in peer valuations changes the appetite. The answer was a restatement of policy rather than a signal, and it included an unprompted explanation of what the company's equity toehold positions are actually for.

Q: "And then, can I ask on M&A? I mean, particularly given the pullback in valuations and a bit of a derating in the sector, and on the back of the closing of the Finnish acquisition, and considering, I mean, the dozens of current toehold equity positions that you guys have, how is Agnico now viewing the potential for further acquisitions?"
— Lawson Winder, Bank of America

A: "We have never had direction from the board to get bigger just for the sake of getting bigger. So it is our job. We get paid to look at opportunities to wisely invest our owners' money. That means in projects. That means we look at exploration. That means we do look at M&A opportunities all the time. And our toehold investments, I'll just say it again, it's not to have a portfolio of assets. It's really items that we might be interested in so that we can learn more about them."
— Ammar Al-Joundi, President and CEO

Assessment: The answer is boilerplate on intent and informative on mechanics. The toehold portfolio is explicitly an information asset, not an investment portfolio, which sits awkwardly beside $261 million of it being sold this quarter to fund buybacks and a $155.3 million gain being booked. Both things can be true, but a company that sells its "learning positions" when it wants repurchase capacity is running a portfolio whether it calls it one or not.

What They're NOT Saying

1. What Barnat does to 2027 and 2028 costs. The production effect was quantified within 24 hours of the event and repeated in the release. The cost effect was acknowledged in direction only, deferred to a multi-year budgeting process that has just begun. Losing up to 150,000 low-cost open-pit ounces a year for two years, at a mine already guided to $1,260 per ounce for 2026, has a mix effect on the consolidated cost line that is calculable from the company's own mine plan. Nobody has published it.
2. How the sterilised ounces are treated in the reserve statement. The operating executive said explicitly that he will not book their recovery and that February will bring a better view. That is an honest answer and it leaves open whether 370,000 ounces come out of reserves at year end, which matters for the reserve-per-share metric the company itself emphasises.
3. The carrying value of the contingent value rights. 207,654,166 CVRs are outstanding, each worth up to $3.00 in cash across three reserve and production milestones over a ten-year term. Maximum aggregate exposure is roughly $623 million. The release states the count, the term and the milestones, and does not state what was recognised on the balance sheet or how the liability will be remeasured.
4. The ounce cost of the Canadian Malartic safety shutdown. The six-day unscheduled mill shutdown related to the April fatality is named as a driver of that mine's production shortfall. It is not quantified, which makes it impossible to separate the safety cost from the throughput and grade decline in the same sentence.
5. Any target or timeline for Kittila mill recovery. Recovery has now been disclosed as "lower than planned at 81%" for two consecutive quarters, with four named improvement initiatives and no target recovery rate and no date. The word does not appear anywhere in the call transcript, so this is a disclosure the company makes and nobody follows up on.
6. Anything about buyback price discipline. The payout policy is a percentage of free cash flow. It contains no reference to price, and no question on the call asked whether $178.86 was a good average against a stock that closed the quarter at $155.13. A company that repeatedly frames its strategy around per-share value creation has a repurchase policy that is indifferent to the share price.
7. A free cash flow outlook of any kind. Every other major line has a guidance range. Free cash flow, the metric the shareholder-return policy is denominated in, has none, which means the 40% payout target has an undisclosed denominator. Management noted the potential to exceed the target for the full year without attaching a number to either side of the ratio.
8. What "near the lower end" means numerically. The range is 3.3 to 3.5 million ounces and the release says production will be near the lower end. Whether that is 3.30, 3.35 or 3.40 million ounces is a 100,000-ounce spread, worth roughly $400 million of revenue at the current gold price, and it is the difference between comfortably holding the cost guide and testing it.

Market Reaction

• Pre-print setup: AEM closed at $144.43 on July 29, down 14.8% year to date from $169.53, down 6.1% over the trailing 30 days from $153.76, and up 14.4% over the trailing twelve months from $126.30. The 52-week closing range entering the print was $123.37 to $252.19. The S&P 500 was up 6.9% year to date over the same stretch.
• The stretch since our last note: AEM fell 21.3% from the $183.56 close on May 1. Gold spot fell 13.1% from $4,644.50 to $4,036.30 over the same period, and the VanEck Gold Miners ETF fell 15.5% from $87.11 to $73.57. AEM underperformed its own sector proxy by roughly 580 basis points.
• Barnat disclosure day: on July 2, AEM closed down 0.6% at $153.86 while gold rose 1.1% and the gold-miner ETF rose 4.5% to $78.43. That single session is roughly 510 basis points of company-specific underperformance and is the cleanest available read on what the market charged for the wall movement.
• After-hours move: shares rose 2.73% to $148.37 in the after-hours session following the July 29 release.
• Next-day session: AEM opened at $146.63, a 1.5% gap, traded between $144.47 and $150.86, and closed at $150.75, up 4.4% or $6.32. Volume of 2.7 million shares was 1.0 times the 30-day average. The S&P 500 rose 1.7%.
• Peer and commodity context on the same session: gold spot rose 3.1% to $4,160.60, the gold-miner ETF rose 4.4% to $76.78, and Newmont rose 4.8% to $95.76.

The reaction looks like a vindication and is not one. AEM's 4.4% gain matched the gold-miner ETF's 4.4% to within a basis point and lagged the largest North American peer, on a session where the metal itself rose 3.1%. On volume that was exactly average. A print that beat on production, came in below both cost midpoints and delivered a record buyback bought precisely nothing in relative terms. The market treated Agnico as a gold proxy on the day it reported its best operating quarter in a year.

That is the correct interpretation of the whole quarter, not just the session. Between May 1 and July 29 the company sanctioned its largest project, closed a 2,492 square kilometre land consolidation, quadrupled its buyback, added $352 million of net cash and printed unit costs below both guidance midpoints. Over the same window the equity lost a fifth of its value, and it lost more than the sector because a pit wall moved. Nothing management did in the quarter registered against a $608 per ounce move in the price of gold.

The pre-print setup explains the muted relative response. Entering the release the stock had already de-rated 21.3% from our last note and sat 42.7% below its 52-week closing high, so the positioning question was whether the print contained a fresh negative rather than whether it contained a positive. It did not, and the stock traded with its commodity. The reaction to watch was never the day; it is whether the second consecutive quarter of production above budget starts to compress the discount at which this equity trades to spot gold. So far it has not moved at all.

Street Perspective

Debate: Is Barnat a Three-Year Hole or a Contained Event?

Bull view: The bull case being made on the Street is that 370,000 ounces across three years is roughly 3.7% of the production base, that Barnat was an end-of-life pit whose closure was already scheduled, that no one was hurt and no equipment was damaged, and that the monitoring systems worked exactly as designed. The full-year range held and the consolidated cost guide held with it.

Bear view: The bear camp contends that the ounces removed are cheap open-pit ounces from the company's second-largest Quebec asset, that the 2027 and 2028 cost effect is unquantified precisely because it is unattractive, and that a geotechnical failure on a monitored wall at a mine the company has operated for over a decade is not a random event. Enhanced monitoring detected movement and the wall still failed.

Our take: The bulls have the arithmetic and the bears have the sequencing. The production number is genuinely small against the base. The cost number is not small at Malartic, and it lands in years for which no guidance exists and against which the offsets named this quarter (a strong dollar, low royalties, cheap by-product assumptions) may have reversed. February is when this gets settled, and February is also when the reserve treatment lands.

Debate: Does the Capital Program Still Fund Itself at $4,000 Gold?

Bull view: Some sell-side desks argue the answer is obviously yes: net cash rose $352 million in a quarter that included $801 million of capital, $578 million of Finnish acquisitions and $625 million of shareholder returns, with the revolver undrawn and roughly $2 billion of liquidity untouched. The tax burden is 70% behind for the year, which frees roughly $1.3 billion of second-half cash flow relative to the first half.

Bear view: The bear camp points out that free cash flow before working capital, the clean measure, was $1,303 million in a quarter with a $4,483 realized price, and that spot is now 10% lower while capital guidance just rose $415 million. Run the second half at spot with the higher capital number and the free cash flow that funds a 40% payout, a growing dividend and a $2 billion repurchase ceiling gets materially thinner.

Our take: Both are right about different years. 2026 is funded and then some, because the tax profile front-loaded. 2027 is where the question bites: Hope Bay construction is a multi-year spend at its heaviest, Barnat removes ounces, and there is no offsetting one-time tax reversal. The balance sheet absorbs it. The buyback is what flexes.

Debate: Is the Cost Improvement Real or Currency?

Bull view: A growing consensus view is that the second quarter validates the productivity argument: record mill throughput at mines representing 52% of production, automated haulage productivity up 65% at one operation, development rates up 14% at another, and a three-year track record of offsetting roughly half of input inflation. Kittila held unit costs flat year over year through the worst of the sector's inflation.

Bear view: The skeptics note that the by-product credit inside total cash costs rose $29 per ounce year over year, which is nearly twice the $16 cushion below the guidance midpoint, and that management itself attributed the cost performance to a stronger dollar, lower royalties on a lower gold price, and conservative copper and silver assumptions. Strip those and the quarter is in line, not ahead.

Our take: Roughly half and half, and the company said so without being pushed, which is worth something. The productivity evidence is specific and verifiable at the mine level, which is more than most of this sector offers. The honest framing is that Agnico can offset about half of input inflation through operations and needs currency or by-products for the rest, and that this quarter it got both.

Debate: Does a 21% De-Rating Create an Entry Point?

Bull view: The bull case is that at $144.43 the equity discounts a forward realized gold price near $3,695 per ounce against a $4,036 spot and a $4,500 assumption inside the company's own cost guidance, on an asset base that is 85% Canadian, carries $3.3 billion of net cash, funds a decade of growth from cash flow, and just produced above budget for a second consecutive quarter.

Bear view: The bear camp contends that this is the same argument made three months ago at $183.56, that it did not work, and that the discount to spot has stayed constant while spot fell 13%. If the market will not close the gap on good operational news, the gap is not a mispricing, it is the market's estimate of the through-cycle gold price. Gold equities carry roughly three times the metal's beta and the metal's trend is down.

Our take: The bear argument is the stronger one about the last three months and the weaker one about the next twelve. A discount that stays constant while the underlying halves in value is not evidence the discount is right, it is evidence it is stable. The reason to stay long is not that the gap must close; it is that at $144.43 you are paid for a gold price 8.5% below spot on a business that has now demonstrated it can hold its cost guide, which is a different proposition from the one at $183.56.

Model Update & Valuation Framework

Changes to Our May Estimates

FY2026E Assumptions

Two assumptions carry most of the risk, and they are different ones from May. The first is the second-half gold price, held flat at the July 29 spot rather than at any forecast; a $100 per ounce move is worth roughly $160M of EBITDA over the remaining half. The second is the second-half cost path, where we model $1,083 per ounce against Q2's $1,054, because the diesel purchased 30% above budget, the Malartic step-up toward $1,260 and the scheduled shutdowns at six mines all land after June 30. The cost cushion below the guidance midpoint is real and it is roughly one quarter's worth of the diesel overrun.

Valuation and Price Target

At the July 29 close of $144.43 on 506.4 million shares outstanding net of shares held in trust, the market capitalisation is roughly $73.1B. Deducting $3,267M of net cash gives an enterprise value of roughly $69.9B, or 6.8 times our FY2026E adjusted EBITDA of $10,330M. That compares with 7.6 times on the equivalent basis when we initiated in May. On our estimates the free cash flow yield is 5.3% and the adjusted price-to-earnings ratio is 12.8 times.

We cut the 12-month price target to $175 from $210, roughly 21% above the pre-print close. The multiple is unchanged at 8.7 times enterprise value to adjusted EBITDA plus net cash, which we continue to regard as a defensible premium to the sector for jurisdictional concentration, net cash and a funded growth profile. The target sits between two anchors we consider equally legitimate. Applying 8.7 times to FY2026E adjusted EBITDA, which contains two quarters of realized prices that will not repeat, produces $184. Applying the same multiple to the second half annualised at the current spot price produces $164. We set the target in the lower half of that band because seven months of the year are behind us and the forward number is the one a 12-month target should be anchored to.

Gold Price Sensitivity

The table below runs a forward-12-month version of the framework rather than the calendar-2026 version, because the first half is already banked and tells you nothing about the next four quarters. It holds production at 3,300 thousand ounces, which allows for the Barnat effect in 2027, and the multiple at 8.7 times. Royalty and NTI costs are assumed to absorb roughly 2.5% of the change in revenue in each direction, so the cost line moves with the price.

Thesis Scorecard Post-Earnings

The table scores the same six pillars established in May. Two status tags moved this quarter, in opposite directions.

Overall: Thesis intact and repriced. The company-specific case is in better shape than in May: the cost pillar that was deteriorating has stabilised, the balance sheet strengthened through a heavy spending quarter, and every dated commitment from the last call was met. The two risks that have worsened are the commodity, which is outside management's control, and safety, which is not. Net of both, the thesis is where it was and the stock is 21% lower.

Action: Hold existing positions and add on weakness, sized for commodity volatility. The entry price discounts a gold environment roughly 8.5% worse than the one that exists, which is the same margin of safety we identified in May at a price 21% higher. The difference now is that the downside scenario has been demonstrated rather than described: a $600 per ounce move in gold took a fifth off this equity in three months while the business improved. Position accordingly, and treat a sustained break of the $3,700 per ounce level in spot gold as the point at which the framework stops supporting the rating.

Commitments to Watch Next Quarter

1. Barnat remediation completing in the third quarter and mining resuming in the fourth. Berms of 15 to 25 metres, a redesigned pit and a final design still being refined. Any slip past Q4 puts the "near the lower end" production framing at risk rather than the midpoint.
2. Second-half total cash costs at or below roughly $1,067 per ounce if the $1,070 full-year midpoint is to hold, against Q2's $1,054 and our own $1,083 estimate. The range itself is not threatened until $1,167.
3. Second-half AISC near $1,479 per ounce to hold the $1,475 midpoint, with sustaining capital per ounce already running $286 in Q2 against $265 for the half.
4. Whether the remaining 30% of Nunavut diesel is bought at a better price in the third quarter than the 30%-above-budget level paid for the first 130 million litres.
5. The full Rupert share count in the Q3 weighted average, which steps up roughly 0.9% from the 501.7 million that divided this quarter's earnings, and whether the buyback offsets it.
6. Buyback pace and price. A payout target above 40% of free cash flow with a lower free cash flow base, and whether repurchases are executed with any reference to the share price after averaging $178.86 into a $155.13 quarter-end.
7. Kittila mill recovery above 81%, still with no target and no date attached after two quarters of disclosure.
8. Boston deposit assays, expected in the October release, from the first exploration drilling at Hope Bay's second deposit since the 2021 acquisition.
9. Odyssey ramp development reaching 2,000 metres per month in the fourth quarter, against roughly 1,800 currently, and the headframe changeover starting in Q3 without disturbing the Q2 2027 commissioning date.
10. Any quantification of the 2027 and 2028 Canadian Malartic cost effect before the February guidance release, and how the 370,000 sterilised ounces are treated in the year-end reserve statement.

Bottom Line

Agnico Eagle did in the second quarter almost exactly what we said in May it needed to do. Production came in at 855,816 ounces, 3.4% above the compiled estimate, after missing by 4.0% three months earlier. Total cash costs of $1,054 per ounce landed $16 below the full-year guidance midpoint, and AISC of $1,459 landed $16 below its own, after both finished above in the first quarter. The buyback stepped from $150 million to $400 million, which is the number we said to grade against. Hope Bay was sanctioned in May, on schedule, at roughly $2.4 billion. The portfolio sales that management said would fund repurchases funded $261 million of them. Net cash rose $352 million through a quarter that spent $801 million on capital and $578 million on Finland.

Two things went the other way. On July 1 a million tonnes moved on the north wall of the Barnat pit, sterilising roughly 370,000 ounces across three years and pushing Canadian Malartic's full-year unit cost to approximately $1,260. And on May 1, at Upper Beaver, a third employee in twelve months died at work. The first is a mining event with a quantified production consequence and an unquantified cost consequence. The second is a pattern, and this quarter it showed up in the production numbers for the first time.

Neither explains the stock. Between our last note and this print, gold fell $608 per ounce and Agnico Eagle fell $39.13 per share, a 21.3% decline against a 13.1% move in the metal and a 15.5% move in the sector proxy. On the day it reported, the stock rose 4.4% and so did the gold-miner ETF, to the basis point, on exactly average volume. The market is trading this security as a gold instrument and is not currently paying for operational execution, which is either an opportunity or a correct assessment of what a gold miner is.

We maintain Outperform and cut the target to $175. The May target was wrong by 45% against the price at which this print landed, and it was wrong for one reason: it assumed $4,700 per ounce gold. At $144.43 the equity discounts a forward realized price near $3,695, an 8.5% haircut to spot, which is the identical discount we measured in May at a share price 21% higher. Nothing about the company's standing in that calculation has deteriorated. The balance sheet is stronger, the cost line is better, and the pipeline is larger and now funded through its largest commitment. What has deteriorated is the price of the thing it sells, and buying this equity remains a bet that gold does not keep falling. That bet is now being made 21% lower, on a business that has spent the quarter proving it can hold its guidance, which is the only reason to make it twice.