And just like that “The Week Ahead” turns into the weeks ahead (this is our last newsletter of 2025) and then it’s the year ahead! Business newsflow might be light this week, but that may well give room to explore themes, not least the notion of the AI Bubble, to which we try lend some hard facts, see below. And it’s time for what Stock Trader’s Almanac’s Yale Hirsch called the Santa Claus Rally (the last five days of the year and first two trading days of the next year.)

Let’s go!

The full archive and Epistrophy research database is live here: https://epistrophy.beehiiv.com (your email address works as a password). Check it out.

As always, I’m focused on three things:
1) Technology-driven change;
2) the latest in innovation and startup trends, and;
3) stock fraud.

Companies Discussed

In This Note:

• Companies Discussed

• Talking Heads

• Tweet O’ The Week

• Epistrophy In The News

• 📆 of Epistrophy Events

• Availability This Week

Talking Heads

Recent commentary from AI CEOs suggest the AI pessimism doesn’t match reality.

It’s remarkable in recent weeks how weak the Artificial Intelligence sentiment is among investors and the general public, and yet how strong the commentary and results have been from the companies we cover. 

So for our last research piece of the year, we thought we’d look at some actual quotes in the last few weeks from AI executives from earnings calls and investor conferences, to see what picture emerges (emphasis added). 

The most telling signal isn’t a single forecast or a raised guide. It’s how consistently executives describe demand as durable, broad-based and multi-year. 

Gary Smith, Ciena (CIEN: NASDAQ) CEO, Q4 earnings call, Dec. 11: “We continue to see accelerating demand from our cloud customers, including large hyperscalers and emerging neoscalers. Orders from cloud providers are very strong and ramping across our portfolio, constituting a substantial portion of our growing backlog.”

Hock Tan, Broadcom (AVGO: NASDAQ) CEO, Q4 earnings call, Dec. 11: “We see the spending momentum by our customers in AI continuing to accelerate in 2026. AI revenue grew 65% year-over-year to $20 billion, and this growth trajectory now exceeds tenfold over the eleven quarters we’ve reported this business.”

Gary Smith: “The simple truth is that AI continues to drive network expansion across all our customer segments, and the scale of investment currently underway is massive and accelerating faster than anything we, or indeed the industry, have seen to date.”

In December, executives weren’t just speculating about future projects. They were describing deployments already constrained by logistics, lead times, and physical capacity.

Jeff Clarke, Dell Technologies (DELL: NYSE) COO, UBS Global Technology and AI Conference, Dec. 2: “We added almost $7 billion to our backlog, bringing it to $18.4 billion, driven by record AI orders. Customers are deploying thousands of GPUs at scale, and the cycle time from design to installation is compressing as demand continues to strengthen across hyperscaler, neocloud, and enterprise environments.”

Trent McKenna, Comfort Systems USA (FIX: NYSE) COO, Sidoti Investor Conference, Dec. 11: “We’re seeing a really robust pipeline looking forward into the next several quarters, particularly in data centers and chip manufacturing. These projects are episodic and programmatic, but the volume and consistency of opportunities point to a sustained buildout environment.”

Jeff Clarke: “To deploy thousands and thousands of GPUs in a short period of time requires upfront engineering, power optimization, cooling density and rapid installation. What we’re seeing now is customers prioritizing speed of deployment, because these systems are going straight into revenue-generating environments.”

Across sectors, hyperscalers emerge as the gravitational force. Their capital intensity sets the pace not only for servers and silicon, but for networks, memory and the trades required to wire everything together.

Gary Smith : “Cloud providers are as focused on scaling their networks as they are on access to power. They cannot and do not intend to strand their significant investments in AI-related data center infrastructure, which means network expansion is a prerequisite, not a discretionary decision.”

Hock Tan: “Demand in AI networking has been even stronger as customers build out data center infrastructure ahead of deploying accelerators. Our current backlog for AI switches exceeds $10 billion, driven by hyperscalers preparing their environments well in advance of compute deployment.”

Ash Kulkarni, Elastic (ESTC: NYSE) CEO, Barclays Global Technology Conference, Dec. 10: “We saw very strong customer commitments this quarter, including some of the largest deals we’ve ever done. Those commitments convert into revenue over the next twelve months, and AI continues to be the fastest-growing part of our business.”

Gary Smith: “These customers have sustainable business models and the capital to invest. What constrains them today is not willingness, but the physical ability to scale networks fast enough to operationalize AI at global scale.”

The numeric expression is backlog. Oracle’s December disclosures, framed as “remaining performance obligations” or RPO, moved the AI debate from quarterly volatility to multi-year visibility — and it became the bizzare nexus of AI doubt.

Doug Kehring, Oracle (ORCL: NYSE) chief accounting officer, Q2 FY26 earnings call, Dec. 10: “Remaining performance obligations ended the quarter at $523.3 billion, up 433% from last year, driven by contracts signed with Meta, NVIDIA, and others. The vast majority of these bookings relate to opportunities where we have near-term capacity available.”

Clay Magouyrk, Oracle CEO: “We continue to experience significant and unprecedented demand for our cloud services, and we are increasing capital expenditures to meet that demand – where it meets our profitability requirements and where capital is available on favorable terms.”

Larry Ellison, Oracle Chairman & CTO: “Demand for Oracle Cloud Infrastructure is unprecedented. Customers are signing long-term contracts for massive GPU capacity because they cannot get this scale elsewhere, and they need it now to support AI workloads.”

Doug Kehring: “We now expect fiscal 2026 CapEx will be approximately $15 billion higher than previously forecasted, reflecting the pace at which committed customer backlog can be converted into revenue.”

Semiconductors tend to surface stress before optimism. When memory suppliers start talking about multi-year supply gaps and pervasive demand, it’s usually because orders are already exceeding physical output.

Sumit Sadana, Micron Technology (MU: NASDAQ) Chief Business Officer, Q1 FY26 earnings call, Dec. 17: “There has been a very significant and pervasive step-up in demand across the data center customer base. This mismatch between supply and demand continues for the foreseeable future, and it is affecting HBM, conventional DRAM, and NAND across all segments.”

Mark Murphy, Micron CFO: “This is a multi-year build-out driven by well-capitalized customers. AI requires more and better memory, and the supply gap is not easy to address in the near term given the capital intensity and complexity of expanding capacity.”

Sassine Ghazi, Synopsys (SNPS: NASDAQ) CEO, Q4 earnings call, Dec. 10: “We are operating amidst a multi-trillion-dollar AI infrastructure build-out, driving robust semiconductor demand and design starts. AI is increasing engineering complexity while accelerating the pace at which new systems must be designed and deployed.”

Sumit Sadana: “Our mix will increasingly tilt toward data center workloads, but the broader point is that demand has stepped up faster than industry supply can respond, and that dynamic underpins our confidence looking forward.”

Taken together, these voices describe the same system from different angles: networks strained by traffic that must be monetized, data centers rising faster than labor can wire them, memory fabs stretched by unprecedented intensity, and balance sheets reshaped by contracted demand.

Tweet O’ The Week

Epistrophy In The News

Yahoo! News, quoted me once again explaining why Oracle has become the poster child for AI bubble. A CEO friend said to me after that hit afterwards: “so basically you’re the ‘only Oracle fan-boy”? Low key...

On Yahoo! Finance, I discussed rising risk at OpenAI and why Wall Street continues to misread capital spending disclosures at Oracle. The problem is not that Oracle is spending too much. It is that investors see reflexive skepticism where clarity is actually increasing.
https://x.com/YahooFinance/status/2000673231204659202

And with Connell McShane of NewsNation, we unpacked a new report from Vanguard arguing that AI adoption may increase hiring over time rather than reduce it. The nuance matters.

📆 of Epistrophy Events

Availability This Week

Holidays or not, I’m around! I’m in the San Francisco office all week or at least can be with the drop of a hat. Feel free to text for time-sensitive items; email works too, replies later in the day.

Written reports are available to clients, with video summaries on YouTube, and of course our popular summaries of the summaries on Instagram, TikTok, and YouTube Shorts.