Akamai: Build for Anthropic—or Preserve Room to Maneuver?
Download case PDF (6 pages)Request teacher’s pack
Teaching case
Information cutoff: October 3, 2026
Audience: MBA students and working professionals
A contract changes the scale of the question
On September 24, 2026, Akamai Technologies announced that Anthropic had committed to purchase $11.6 billion of cloud services over seven years. The agreement supports Anthropic's growing CPU workload needs on Akamai's distributed cloud infrastructure. The announcement also described a possible additional $9 billion of purchases during the agreement term, subject to mutually agreed terms. That option brought the total potential commitment to approximately $20 billion, but the additional purchases were not contracted demand.
For Akamai, the agreement was both a customer win and a capacity decision. The company estimated about $5.5 billion in capital expenditure related to the base commitment. It expected to spend approximately $1.7 billion more in 2026 to secure and pre-purchase supply-chain components, including memory. The timing mattered: the company forecast that service would begin in late second quarter 2027, with revenue ramping during the second half of that year and the full contracted revenue run-rate reached by the end of 2028.
The public announcement did not describe a board vote on how to treat the possible expansion. For classroom discussion, the decision is framed prospectively: as of October 3, Akamai's leadership must decide what capital-allocation rules and capacity posture to use while carrying out the base contract and preserving the possibility of negotiating additional purchases.
Akamai's starting position
Akamai had built its business around internet delivery, security and, increasingly, cloud infrastructure. Management described a network spanning thousands of points of presence, offering compute from core to edge on diversified hardware. Its strategic claim was that the footprint and experience serving large enterprises could support fast, reliable and secure applications and workloads.
The company's second-quarter results showed that the cloud infrastructure business was growing from a smaller base. Akamai began reporting cloud infrastructure services (CIS) separately in 2026, describing it as a primary growth area and a significant investment focus. Security remained the largest of the three reported solution categories.
Exhibit 1. Akamai revenue by solution
| $ millions | Q2 2026 | Q2 2025, recast | YoY change reported by Akamai |
|---|---|---|---|
| Security | 604.436 | 551.914 | 10% |
| Delivery and other cloud applications | 395.927 | 420.117 | (6)% |
| Cloud infrastructure services | 99.319 | 71.463 | 39% |
| Total | 1,099.682 | 1,043.494 | 5% |
Source: Akamai 2026 Q2 Form 10-Q, Note 3, “Revenue by Solution,” and MD&A, “Revenue by Solution,” printed pp. 23 and 29–30. Dollar values are reported in $000 and shown here in $ millions. Prior-year amounts were recast after CIS became a separately reported category. Total and overall growth are calculated from the reported values; overall growth is rounded.
Management attributed year-over-year growth primarily to security and CIS, partly offset by lower delivery and other cloud applications revenue. It attributed that decline to lower pricing on contract renewals. In the filing, Akamai also cited higher colocation costs, bandwidth and network-equipment depreciation as it expanded the platform, and significant increases in server and memory costs related to market dynamics driven by hyperscalers.
The contract's economics—and what remains unknown
The September investor presentation provided more detail than the press release. It showed management's estimate of $5.5 billion in total capital expenditure for the base commitment, with a front-loaded supply purchase followed by build-out. It also forecast the revenue ramp. In the financial overview, Akamai characterized the multi-year deal economics as producing “strong cash conversion and operating margins.” The materials offered no definition or calculation bridge for “strong,” no contract-level operating-cost schedule, and no cash receipt schedule.
The contract also included an equity warrant. Akamai said that about 3.1 million shares, approximately 2% of shares outstanding at the time, were expected to vest with the base commercial arrangement. Additional warrant shares would vest in increments tied to further purchases, up to approximately 5% in total if the relationship expanded to the stated potential amount. The presentation said the fair value of shares expected to vest with the base arrangement reduced contract revenue; additional dilution would depend on share price and any further commitments.
Exhibit 2. Announced base commitment: public estimates and terms
| Item | Disclosed amount or timing | How to interpret it |
|---|---|---|
| Base contractual commitment | $11.6bn over seven years | Announced purchases; not revenue already recognized or cash already collected |
| Potential further purchases | Up to $9bn, by mutual agreement | Optional expansion; not part of the base contract |
| Estimated base-related capex | Approximately $5.5bn | Management estimate; not the total cost to serve |
| Additional 2026 capex | Approximately $1.7bn to secure/pre-purchase supply components | Management estimate included in the investment plan |
| Revenue ramp | $150–300m shown for FY2027; full contracted run-rate expected by end-2028; about $1.7bn annually thereafter | Company's revenue estimate; not a cash collection schedule |
| Capacity/power | Approximately 95–105 MW to deliver previous large commitments plus the new contract | Portfolio estimate combining prior and new commitments |
| Revenue per MW | Approximately $22m annually, as presented by Akamai | Revenue only; not contribution margin or free cash flow |
| Warrant | About 2% expected to vest with base arrangement; about 1% for each further $3bn purchase tranche; up to about 5% | Subject to vesting terms; not all shares necessarily vest at announcement |
Source: Akamai's Sept. 24, 2026 investor presentation, slides 3–7 (PDF pp. 2–6), and Exhibit 99.1 to the Sept. 24 Form 8-K. Estimates are stated as of Sept. 24 and subject to change. The 95–105 MW and $22m/MW figures cover prior commitments plus the Anthropic base commitment.
The public disclosures did not provide the contract's cost of hardware, power, colocation, bandwidth, support, maintenance or customer-specific operations. They did not show payment timing, working-capital requirements, termination protections, expected utilization or the amount of incremental investment required for optional purchases. Nor did they show how much capacity could be redeployed after the contract term. The stated total contract value and estimated capital expenditures therefore did not, on their own, establish a project return.
Competition for capacity and investment
The cloud market rewarded scale, but scale required long-lived infrastructure and access to power, servers and components. The OECD's 2025 competition review described high fixed and sunk costs, ongoing energy needs, economies of scale and chip-access constraints as barriers to expansion. The review also noted that established cloud platforms benefit from integration with other digital services and from customer familiarity. These market-level findings did not answer whether Akamai's specific investment would be attractive.
The largest providers continued to invest heavily. Amazon reported AWS sales growth of 37% in the second quarter of 2026. It also reported $96.3 billion in cash capital expenditure during the first half of 2026, primarily for technology infrastructure, most of which supported AWS, as well as capacity for other Amazon businesses. Amazon reported AWS operating income of $16.6 billion for Q2, with infrastructure spending partly offsetting the contribution from higher sales.
Microsoft reported $214.4 billion in fiscal-2026 Microsoft Cloud revenue and 41% growth in Azure and other cloud services. It also reported a 66% Microsoft Cloud gross-margin percentage, down year over year as AI infrastructure investment and higher AI product usage affected the measure. Microsoft Cloud includes Azure but also Microsoft 365, LinkedIn commercial services and Dynamics 365. These figures show the breadth and scale of a peer's platform, but they are not directly comparable with Akamai CIS or the Anthropic contract.
These providers' filings offered little basis for a direct comparison of prices, workload mix, switching costs or margins on CPU services. Akamai's advantage could lie in its distributed network, enterprise relationships and security capabilities. Yet neither the contract announcement nor the Q2 filing quantified a security premium, cross-selling benefit, or lower cost to serve Anthropic workloads because of that footprint.
What should Akamai optimize?
The immediate task was to deliver the base commitment as planned. The strategic question was how to preserve room to pursue more. Reserving capacity and buying components early could improve Akamai's ability to meet customer demand, but would use capital before the optional purchases were agreed. A tighter posture could protect liquidity and leave room to invest in security and other CIS opportunities, but might sacrifice supply access or expansion opportunity. Management's margin and cash-conversion claim offered a positive signal, while the lack of a supporting bridge limited outsiders' ability to evaluate it.
The choice was not only how much to invest; it was also how to make future investment decisions. Could Akamai define observable triggers before adding capacity? How should it weigh contract demand against uncertain costs, other customers, market position and potential dilution? What evidence would management need before agreeing to each further purchase tranche?
Decision challenge
As an adviser to Akamai's leadership team, recommend a capital-allocation policy for the period through the base contract's build-out. Should the company prepare aggressively for the potential additional purchases, limit investment to the announced commitment, or adopt another approach? Specify the conditions that should trigger or prevent further investment, the evidence needed to apply them, and how Akamai should balance Anthropic's demand against its broader cloud and security strategy.
Discussion questions
- What does the Anthropic agreement change about Akamai's strategic position? What does it leave unresolved?
- How should the company assess management's “strong cash conversion and operating margins” statement when the supporting economics are not disclosed?
- Which capabilities could distinguish Akamai's cloud offering from larger platforms, and what evidence would be needed to show that these capabilities create economic value?
- What decision rules should govern capacity reserved for optional purchases? How should those rules account for customer concentration, supply availability and dilution?
- Which key metrics should management disclose internally to compare the base contract, optional expansion and investment in other customers or security offerings?
Sources
- Akamai Technologies, Form 10-Q for quarter ended June 30, 2026, filed Aug. 7, 2026, SEC filing, Note 3 and MD&A, printed pp. 23, 29–32, 42, 54–55.
- Akamai Technologies, Sept. 24, 2026, Exhibit 99.1, “Akamai Announces $11.6 Billion Multi-year Agreement with Anthropic,” SEC exhibit.
- Akamai Technologies, “Expanded Strategic Relationship with Anthropic,” Sept. 24, 2026, investor presentation, slides 3–7.
- OECD, Competition in the Provision of Cloud Computing Services (2025), report, pp. 20–25.
- Amazon.com, Inc., Form 10-Q for quarter ended June 30, 2026, SEC filing, MD&A, printed pp. 27, 30–32.
- Microsoft Corporation, Form 10-K for year ended June 30, 2026, SEC filing, Item 1; Item 7, “Intelligent Cloud” and “Microsoft Cloud”; Note 19, printed p. 85.
This teaching case is based on public information available through October 3, 2026. The decision prompt is constructed for classroom use and does not imply that Akamai had formally announced this specific decision as pending.