IBM closed at $245 on 6 November, 26% below the $332 it touched on 2 June, the day after Barclays started coverage at $350 and the company pledged $10 billion to quantum computing145. In between came 14 July, when a pre-announced quarter of slipped deals cut a quarter off the stock in a single session3. The market has since done what it does after a shock: it has gone back to pricing IBM as a slow IT services company with a dividend. We think that misses two things that are both true. The software inside IBM is a compounder, and the quantum program inside IBM is the only one on a public exchange that has a factory, a wafer foundry and a dated fault-tolerant machine. The bears price the second one at nothing. That is where the trade is.
This is the first piece in our equities section, opened in November 2026 around one frame: the next cycle in markets is being built in silicon and in qubits, and the companies that make the picks, the shovels and the machines are where the record will be written. The analysis is as of 9 November 2026.
Where we are now
The year has been a round trip in the wrong direction. IBM entered 2026 at $296 after three straight years of gains of 22%, 39% and 38%14. It spiked to $332 in June on the quantum pledge, and then on 14 July Arvind Krishna pre-announced a second quarter with revenue up 1%, infrastructure down 7% and operating earnings of $2.93 against a $3.01 consensus, with the admission that the company "did not adapt and move quickly enough"3. The stock lost about $55 billion of market value that day, HSBC cut it to Reduce with a $191 target, and full-year growth guidance went from "more than 5%" to 4–5%4.
The third quarter, reported on 21 October, read differently. Revenue was $17.1 billion, up 6% in constant currency against a 1% print the quarter before; software grew 9%, Red Hat 11%, the data business 19% with Confluent inside it; infrastructure returned to growth at 5%; operating earnings of $2.95 beat the $2.86 consensus; and free-cash-flow guidance was raised for the second time this year to growth of more than $1 billion, which implies above $15.7 billion for 202612. The market found the one soft number, a 58.5% gross margin against 59.1% expected, and sold the stock 5% after hours15. Year to date IBM is down 17% while the S&P 500 is up about 13%15.
At $245 the stock trades at 18.7 times forward earnings, 17.3 times EBITDA and 15.8 times free cash flow, a 6.3% free-cash-flow yield with a 2.8% dividend on top14. Sell-side targets sit between $174 and $350 with an average near $266; after the quarter Jefferies and Citi were at $280, Barclays at $277, JPMorgan at $2695.
The bull case
1. The compounder underneath
Software is now about 46% of IBM's revenue and most of its profit. It grew 9% in constant currency in the third quarter at a 35% segment margin, up 210 basis points, on $24.9 billion of annual recurring revenue growing 8%12. Red Hat is growing 11%, OpenShift's $2.4 billion of ARR is growing above 30%, and the two large acquisitions, HashiCorp and the $11.6 billion Confluent deal that closed in March, are being absorbed with 2027 named as the synergy year2. The infrastructure cycle that collapsed in the second quarter has turned: the z17 mainframe has delivered more than $1 billion of incremental revenue over six quarters, and the distributed business grew 32%2.
Put that against the multiple. Microsoft trades at 25.6 times forward earnings, Accenture at 17.8, Oracle at 13.6 with negative free cash flow from its AI capex14. IBM at 18.7 times, with the highest nine-month free-cash-flow margin in its reported history, is priced like Accenture. Accenture does not own a quantum computer.
2. Quantum is engineering now, not a slide
We are careful with this word, because every listed quantum company has a roadmap. IBM has things that exist. Nighthawk, a 120-qubit processor with 218 tunable couplers, went live in January; its second revision runs 7,500 gates and 100,000 circuits a second2. On 6 November a Nighthawk-powered Quantum System Two was switched on in Chicago7. The day before, IBM, Oak Ridge and AMD launched an error-correction platform on which IBM has demonstrated 70 logical qubits and 438 hard operations, against a Department of Energy target of 100 logical qubits and 100,000 operations by late 20288. Three preprints in July claimed verified quantum advantage on Heron hardware; they remain candidates, not confirmed results, and we say so12.
The money is real too. Cumulative quantum signings passed $1.1 billion, the Quantum Network has 340 members, 16 enterprise engagements including AT&T and General Motors were signed this year62. The US government picked IBM for the largest of its nine quantum CHIPS awards, $1 billion for the Anderon wafer foundry, matched by $1 billion from IBM10. DARPA moved IBM into the final stage of its benchmarking initiative in October9. A 511,000-square-foot, $2.5 billion facility for Starling, the 200-logical-qubit machine dated 2029, was approved at Poughkeepsie in September11. And IBM bought HRL Laboratories, a silicon spin-qubit group, as a hedge on its own superconducting bet. This is a capital program with a factory, not a pitch deck.
3. The option is priced at zero
HSBC's sum-of-the-parts is the honest way to see it. Its "synthetic IBM" basket of SAP, Accenture, HP and IonQ, priced at the same share price as IBM, implied $23 of 2030 earnings against IBM's $16.59, and the conclusion was Reduce4. Read it the other way: at $245 you are paying for the software and the consulting at peer multiples and receiving the quantum program for nothing. Barclays' $350 case needs only the option to be priced as an option5. The stock has already given back the entire June narrative premium and then some; what is left is the number.
The bear case
Now the argument against, and it is the one that won in July.
Growth is low, and growth-adjusted the stock is not cheap. Four to five percent revenue growth, consulting flat with its margin down 150 basis points in a quarter, the mainframe business down 20% as the cycle ages, earnings growing about 11% a year against 19% for the peer group4. Debt is $63.8 billion after $11.4 billion of acquisitions this year, there has been no buyback since 2020, and the dividend is 58% of earnings1. Oracle trades at 13.6 times forward earnings14. A compounder growing 4% is a bond with a story.
Quantum is a 2028–29 revenue line with 2026 roadmap risk. Kookaburra, the 2026 roadmap deliverable, had not been announced as of 9 November. The July advantage claims were matched within weeks by a classical GPU cluster on the runtime, if not on the verification. Quantinuum has 99.921% two-qubit fidelity and a 100-logical-qubit machine dated 2027; Google and Microsoft are racing the same 2029 date13. IBM has committed $10 billion, a $2.5 billion building and a $1 billion foundry against zero disclosed quantum revenue, and its CEO says the line becomes material in 2028 or 2029. Options expire.
July showed how fragile the premium is. One quarter of deals slipping and client budgets diverted to AI servers cost 25% in a day. IBM stopped disclosing its generative-AI book of business this year, the metric that carried the 2024–25 rerating15. The January guidance for 2027 has to carry a Confluent synergy story and a Red Hat growth rate management declined to guide. If AI is eating IT budgets, IBM's consulting arm is on the menu.
The July crash reset IBM from a narrative stock to a numbers stock, and the numbers are better than the multiple: software compounding at 9% on 35% margins, free cash flow above $15.7 billion, a 6.3% yield on it. On top sits the one quantum program that has moved from roadmap to real estate, and the bears price it at zero. We do not need the option to pay in this horizon. We need the market to stop pricing it as a liability.
Our view: IBM trades at $300, in a $290–310 band, by the end of Q4 2027. That is 22% above today and still below the June high. The path runs through Quantum Day on 10 December, the January guidance for 2027 and the first quarters of Confluent synergy. Conviction is moderate — about 60% — and we say why: the growth rate is real, the debt is real, and one more quarter like the second quarter would prove the bears right about the fragility.
Our bet runs one way: up, on the numbers, with the option as the kicker. We are not betting on a quantum breakthrough inside the horizon. If IBM closes a week below $200 first, we were wrong — see below.
What would prove us wrong
One direction, one closing level, and the fundamentals that would hollow it out.
$200 is the floor of the 52-week range, set after the July crash. A week closing beneath it says the market has decided the second quarter was the trend and the third was the exception, and the thesis is closed against us, whatever happens afterwards.
Any of these weakens the thesis while the price holds, and we will log it as it happens: January guidance for 2027 below 4% constant-currency revenue growth or below $16 billion of free cash flow; software growth falling under 7% for a quarter; or Kookaburra slipping out of the roadmap without a replacement date, which would mean the 2029 Starling date is no longer credible.
IBM trading in the $290–310 band on the NYSE by 31 December 2027. If neither the target nor the invalidation resolves by then, the thesis is booked as it stands — unresolved and closed. It does not get extended.
When a condition triggers, we book the result and publish the piece. As always.
This is analysis, not investment advice. Do your own research. The thesis is not changed after the fact.