Turkey's 5G auction was a fiscal success. Turkcell's third-quarter results, published on 9 November 2026, show what that success costs the operators who paid for it. They also show where the consumer benefit is starting to appear.
What the numbers show
Turkcell's Q3 2026 results release, filed with the SEC, reports depreciation and amortization of TRY 22,621 million. That is up 10.3% from TRY 20,506 million a year earlier, a rise of roughly TRY 2.1 billion. EBIT margin fell to 16.6% from 17.4%. The company says higher D&A reflects 5G investments and the licence, which it puts at USD 1.2 billion excluding VAT.
The filing also ties currency pressure to the same investment. It attributes financial costs to "FX losses associated with a year-on-year increase in the short net FX position due to 5G investments." In plain terms, a lira-earning operator is carrying a dollar-denominated spectrum obligation. Turkcell's first installment, including the Wireless Usage Fee, was USD 653 million including VAT.
On the revenue side, mobile data use reached 27.8 GB per user per month, up 33% year on year. Blended mobile ARPU was TRY 436.6, down 3.9% in nominal terms. The 5G effect so far is more traffic, not more revenue per user.
The clearest commercial signal is fixed wireless. Turkcell's 5G-based Superbox added 101,000 subscribers in the quarter, its strongest since Q4 2019, and reached 919,400 subscribers, up 35.6% year on year. The company reports a 75% share of the fixed-wireless segment. Its footnote says that share is calculated on the regulator's definition and is as of Q2 2026.
The case for the design, stated fairly
The strongest argument for Turkey's approach is that it raised money and kept prices flat. The Information and Communication Technologies Authority (BTK) reports that 11 packages across 700 MHz and 3.5 GHz sold for USD 2.945 billion, or USD 3.534 billion including VAT. The minimum value was USD 2.125 billion. Three operators each won spectrum: Turkcell took five packages for USD 1.224 billion, Türk Telekom four for USD 1.094 billion, and Vodafone two for USD 627 million.
BTK says citizens get 5G "without any changes to their existing tariffs." It also requires operators to put 5% of revenues into infrastructure each year from 2029 to 2042. Per-lot prices, reported by Hürriyet Daily News, were tightly clustered. The three 700 MHz lots went for USD 425–429 million each, and the 3.5 GHz lots for USD 186–248 million. That looks like a competitive, non-collusive auction with three viable winners.
A state that wants nationwide 5G, domestic industrial capacity and treasury revenue has defensible reasons to bundle those goals.
Where the bundle strains
The trade-offs are visible in Turkcell's accounts. Three features of the design matter most.
- Up-front dollar pricing. BTK's own page shows a minimum tender value of USD 2.125 billion. Bids landed about 39% above it, excluding VAT. Spectrum is a scarce input, so a high price is not automatically a problem. But a price paid in hard currency by operators that earn in lira turns licence cost into FX risk. The Q3 financing line shows that risk arriving before 5G has produced new revenue.
- Flat tariffs. Holding existing tariffs steady is politically attractive. It also limits how operators can recover the licence, and the nominal ARPU decline fits that limit. Operators are left to recoup the cost through volume, upselling and new products.
- Local-content mandates. BTK says the domestic production rate rises from 45% in the 4.5G period to 60% for 5G. It also requires at least 30% of products to be wholly owned by domestic and national companies, starting at 15% in the first stage, depending on availability. Industrial policy goals can be legitimate. But mandates on network equipment narrow vendor choice, and operators carry the delivery risk if domestic products are late or costlier. That risk sits on top of the licence bill.
The evidence in Turkcell's results does not show the mandates failing. It shows a cost stack of licence, FX exposure, capex and domestic-sourcing duties that is borne almost entirely by operators. Nine-month total capex was TRY 136.5 billion, against TRY 76.7 billion a year earlier. Most of that jump is the licence.
Fixed wireless is the payoff to watch
Superbox suggests where 5G spectrum earns its keep. It competes with fixed broadband without new trenching, and it works where fibre is slow to arrive. Policy should leave that competition room to develop.
That has three implications. First, BTK should keep wholesale and fixed-wireless rules technology-neutral, so that mobile spectrum can compete with fibre without special restrictions. Second, a 75% share in one segment should be watched on competition grounds, but not penalised, because it is also evidence of effective use of the 3.5 GHz and 700 MHz holdings. Third, the 2029 investment obligation should be measured against outcomes such as coverage and speed. If it is only an input quota, it adds cost without adding quality.
The local-content tiers deserve the same treatment. The 30% wholly-owned target should be reviewed against actual product availability and price, and be relaxed where domestic equipment is not yet competitive.
Bottom line
Turkey avoided the worst auction outcomes: a failed sale, a single winner, or a price spike that stranded a bidder. Q3 shows the other side of the ledger. Operators have paid for the spectrum up front and absorbed the FX exposure, while ARPU is flat to down. The returns will come from data volume and fixed-wireless growth. Regulators protect that return best by keeping rules stable, technology-neutral and outcome-based.