Back to News

The Big Three: GE Vernova, IBM, and Microsoft in the AI Infrastructure Era

BusinessTechnologyEconomy

The current trading environment is shaped by several overlapping forces. The conflict involving Iran has opened up what is best understood as a 60-day window of elevated headline risk. Despite this, markets appear to be looking through the conflict — a conclusion supported by the behavior of oil, which is down and remaining down. That dynamic is favorable for equities overall.

A strong corporate earnings picture is in place, but it is about to be tested against the artificial intelligence narrative, with Micron's results serving as a key upcoming checkpoint. The more consequential development, however, comes from the monetary policy side. The new Fed chair, Kevin Warsh, has introduced additional uncertainty since the prior discussion of the interest rate environment. He spoke about his task forces and leaned rhetorically toward the price-stability half of the Fed's dual mandate, shifting emphasis away from the employment side. As market participants digest incoming data and try to gauge how it will affect markets, this policy posture is the single largest unknown.

A specific signal worth flagging is the 2-year Treasury yield, which continues to print new 52-week highs. The 2-year yield characteristically attempts to front-run the Fed funds rate, so its rise indicates that short-term inflation expectations are elevated. What makes the current configuration unusual is that this is happening while the stock market has been rallying — the 2-year yield spiked even as equities climbed. On the present session, the stock market is falling, but the move looks less like broad risk-off behavior and more like a rotation out of the very large, heavily-appreciated technology names, particularly those tied to storage and analog chips.

The overall judgment is that the environment remains healthy. The caveat is uncertainty: while clarity is always preferable, the presence of uncertainty generates volatility, builds term premium, and produces moves across the market.

GE Vernova (GEV)

GE Vernova fell considerably more than the broader market on the session, giving up essentially all of the gains it had accumulated over the prior month. Even so, it retained a year-to-date advance of more than 60%. Part of the day's decline is attributable to the same rotation hitting the market broadly — holders who rode the run-up are taking profits — compounded by developments overnight in South Korea.

The core thesis is that GE Vernova is the leader in the infrastructure underpinning the AI data center buildout, where demand is described as astronomical. What distinguishes the company from its peers on a fundamental basis is twofold: above-average earnings relative to peers, and a higher margin trajectory.

A structural tailwind reinforces this. US infrastructure has not been meaningfully updated in some time, which is why stocks tied to infrastructure are experiencing upward momentum — they benefit from AI-driven electricity demand and these longer-term trends. The longer-term theme here is "electronification."

A concrete catalyst is the headline deal between Microsoft and Chevron, referred to as Project Kirby, involving 2.7 gigawatts. This highlights growing hyperscaler demand, and GE Vernova supplies a majority of the relevant generation equipment. This is treated as another news story that reinforces the longer-term investment thesis rather than a one-off. Because compute is ultimately the bottleneck in the buildout, and because building data centers entails ever-larger electricity needs, GE Vernova ranks as one of the top picks for capturing that demand. Combined with its peer-beating earnings and strong margin trajectory, the fundamental case is compelling.

GE Vernova — Technical Picture

The technical setup mirrors the broader market: notable highs were reached, but a second rally failed to eclipse the initial high, which came in at 1,181.95. After a drop-off, the gap level to watch sits around 1,109; filling that gap would be a notable event. There were also several highs clustered right around 1,088.

To the downside, a smaller gap exists at 944, which would require a further decline to reach. More extreme downside reference points include lows near 855 and a double bottom near 805 — these represent longer-term, more expansive price ranges to consider.

On moving averages, price has fallen below the 5-day weekly EMA (dark blue), which comes in at 1,055.94. The next potential moving-average support is the 21-day EMA (teal) near 1,007. On momentum, the RSI took out its longer-term downward-sloping red trend line, but its shorter-term green upward trend line has been broken; nonetheless, RSI remains above the 50 midline, preserving a bullish tilt. The volume profile shows the stock on the verge of dipping below a node where heavier trading activity sits at 1,040 and above, while the next concentrated area is around 980 to 1,010 (a smaller, less distinct node). On the session, the stock was trading at roughly 1,148.74, down about 7%.

(Note: the segment's spoken price levels and the on-screen quote diverge in places, reflecting live narration over a moving chart.)

IBM

IBM was having a strong session — up about 4.5% — even on a down day for the market, helped by two price-target upgrades that morning from Morgan Stanley and JP Morgan. Morgan Stanley advised investors to be scooping it up, and JP Morgan upgraded the name with a price-target hike.

The conviction in IBM is long-standing — it has been a favored name for a couple of years. IBM sits within the enterprise artificial-intelligence theme. The central question in AI investing is who wins on the enterprise side and who wins on the consumer side. IBM, however, is more than a pure-play AI name. Its consulting arm — one of the very flags that contributed to a sell-off earlier in the year — is actually a strategic asset: because of existing consulting relationships with enterprises, IBM is already in the room when those enterprises make their AI decisions, giving it a structural advantage.

IBM also has exposure to tokenization, a broader theme tied to a striking observation from the South Korea developments: how globalized markets have become. Other markets are now influencing the US, reversing the usual pattern in which the US leads the charge. Tokenization is viewed as part of this shift because it represents broader access to markets. On top of that, IBM has exposure to quantum computing.

Much of the day's upgrade activity is attributed to IBM joining OpenAI's "Daybreak" cyber partner program — an AI-driven cybersecurity partnership. Several such partnerships are occurring, including ones involving Mythos and OpenAI, and these bode well for enterprise adoption.

A noteworthy risk — which also bridges into the Microsoft thesis — comes from a study suggesting that 91% of enterprise executives struggle to understand AI dependencies. This lack of understanding has contributed to backlogs and to deals being slowed, because executives don't fully grasp what those dependencies are. This is framed as an interesting new narrative in the broader AI story, since the ultimate question is one of adoption and where revenue dollars will land. Even so, IBM's key advantage is that it is "in the room," it partners with all the AI models rather than forcing a single choice, and it carries exposure to tokenization, quantum computing, and now cybersecurity. It stands as a top enterprise pick.

IBM — Technical Picture

IBM is down year-to-date but had a strong session even as the broader market — including the Dow — slipped into the red. A notable high point sits at 258, around the level of a prior breakout, and price has managed to hold above that level. The day's push higher filled a smaller gap; to the downside, the next gap would be filled near 253. Such gaps are tracked by traders because they can return to prominence far later and act as support or resistance. Another low near 245 was tested several times and held firm. To the upside there are relative highs just shy of 280, after which there are few clear reference areas. There was a very sharp decline from highs near 332.

The setup is interesting because price is breaking above a very steep downward trend line, with the next potential resistance around 280 based on horizontal price activity. The moving averages are largely clustered together between roughly 258 and 266. While clustered moving averages often indicate sideways price action, that is not the interpretation here; instead, extreme volatility has left the moving averages "confused," unable to establish a clear direction and congregating toward one another. The long-term 251-day EMA (orange), representing one year, comes in near 261 and is the most important one to track; price remains above it. RSI has invalidated its downward-sloping red trend line and is on the verge of crossing back above the 50 midline, which would be a bullish development. The volume profile has two distinct areas: a lower zone containing the point of control (the thick red line marking heaviest trading) spanning 240 to about 260, with the point of control at 258 — price remains above it. After a gulf, the next heavy trading area runs 280 to 310, with nodes around 285, 290, and 305. IBM was trading at roughly 263.73, about 4.5% higher on the morning's two upgrades.

Microsoft (MSFT)

A significant deal out of Microsoft the prior day supported a continued move higher, with the stock up about 1% to 1.5%.

The candid assessment is that the market had been waiting for Microsoft to innovate and compete with what is being seen from Claude (Cowork) and from Codex on the OpenAI side, and Microsoft had not delivered that visibly. The reframing here is that investors need to be more open-minded, and Microsoft's recent deals are evidence of that. The argument loops back to the IBM point — the study that 91% of executives don't understand AI dependencies — but pivots to where enterprise contracts actually reside: with Microsoft, through Microsoft 365 and Copilot. Copilot has surpassed 20 million paid seats, and the enterprise/commercial backlog rose to a record figure cited as 27 billion. That is taken as proof of where enterprise adoption truly is.

A broader, somewhat abstract point about society reinforces this: people are at very different stages of becoming power users of AI. At one extreme, an advanced user might run AI agents that automatically gather information, operable from a phone and wired into a Slack channel shared with coworkers. At the other end, parents might only now be starting to use ChatGPT as a search engine. The adoption of any productivity-enhancing technology — like the internet — takes time, and it takes time with the masses. The crucial question behind the large capital-expenditure spends is where revenue will ultimately translate. If enterprise adoption is genuinely concentrated where Microsoft sits, then it does not matter that Microsoft's products may lack the "coolest bells and whistles" prized by power users — what matters is where the revenue and the commercial backlog are.

For these reasons, the earlier sell-off in Microsoft is judged to have been overblown. The recent deals show more momentum, and Microsoft's broad enterprise distribution channel could meaningfully help the free-cash-flow story. The conclusion is that Microsoft is a buy. After being away from the name for quite some time, the stance is to get back into the position — implementable via a cash-secured put or any preferred method of adding it to a portfolio.

A key context point: Microsoft remains the worst performer in the "Magnificent 7" year-to-date, down more than 20%, even as its fundamentals are described as very strong.

Microsoft — Technical Picture

Despite a notable decline, from a technical standpoint Microsoft did not take out its old lows, which sit near 356.28. There is also a zone of repeated highs that formed a short-term ceiling before a breakout, giving a band of roughly 356 to 374. For now, price seems to have held; there was a dip below that level the prior day, but price is hovering right at it and is also potentially about to break out from a trend line. A push above the trend line combined with a bounce from the support area would be noteworthy.

Additional reference points: short-term relative highs near 381; lows that roughly align with a subsequent high near 401; and longer-term levels at 466 and 478 — the latter significant because it marked a gap after a January earnings event. The 5-day weekly EMA (dark blue) lines up with the trend line, so a break above roughly 378 would be noteworthy from two perspectives at once. The 21-day EMA (teal), representing one month, comes in just shy of 400 — another upside level to watch. RSI is "jumping the gun" ahead of price; importantly, it did not cross into oversold territory, which is more of a bullish sign, and it is starting to break above its downward-sloping trend line. The standard caveat applies: one typically waits for price to act first, since price is always more important than ancillary indicators — but momentum is starting to improve alongside a potential breakout. The volume profile shows heavy trading activity between about 368 and 376 — the area that held firm recently — reinforcing it as important support. The next node sits between about 395 and 425, with the point of control (heaviest trading of all) at 401.90, making the 400 level a standout to the upside. Microsoft was trading around 370, under some pressure but moving higher by about 1%.

Synthesis

The three picks share a common thread: each is a vehicle for capturing the artificial-intelligence buildout, but at different layers of the stack. GE Vernova is the physical-infrastructure and power-generation play, riding electrification and the electricity demand created by compute as the binding bottleneck. IBM is the enterprise-integration and consulting play, positioned "in the room" with existing client relationships and diversified across tokenization, quantum computing, and cybersecurity, while remaining model-agnostic. Microsoft is the enterprise-distribution and adoption play, where the actual paid seats and record commercial backlog reside, making it the place where AI capex is most likely to convert into revenue over time.

Running beneath all three is one recurring insight: the AI story is ultimately about adoption and revenue translation, not flashy features — and the friction created by executives who do not yet understand AI dependencies (the 91% figure) is the key risk shaping how quickly that revenue materializes.

Comments