I created a simple FMCW sweep using MATLAB in baseband, BW = 100MHz:

sweepTime = 1e-3; % 1 ms sweep
BW   = 1e8; %100 MHz;
fs = 2e8;% 200 MHz sample rate

% FMCW waveform: only up sweeps
waveform = phased.FMCWWaveform(...'SweepTime', sweepTime, ...'SweepBandwidth', sweepBW, ...'SweepDirection', 'Up', ...'SweepInterval', 'Positive', 'SampleRate', fs, ...'NumSweeps', 4); % Repeat 4 up-sweeps

% Generate waveform
x = waveform();

% Plot spectrogram
[S, F, T] = spectrogram(x, 1024, 512, 1024, fs);
imagesc(T, fftshift(F - fs/2), fftshift(mag2db(abs(S))))
xlabel('Time (s)')
ylabel('Frequency (Hz)')
title('Spectrogram of 200 MHz FMCW Up-Sweeps')
colorbar

it plotted this

I then try to up-convert it to 2.45GHz,

%checking the sprectrol gram in RF
fc = 2.45e9 ; 

% Extract the first 1 ms sweep (first sweepTime worth of samples)
N_sweep = round(sweepTime * fs); % Number of samples in 1 sweepx1 = x(1:N_sweep);                % First sweep

% Time vector for first sweep
t = (0:N_sweep-1)' / fs;

% Upconvert to RF (use same t as baseband generation!)
x_rf = x1 .* exp(1j * 2 * pi * fc * t);  % complex RF signal

% Windowed FFT for better spectral leakage control
X = fft(x_rf .* hann(N_sweep));
f = (0:N_sweep-1)/N_sweep * fs + fc;

% Plot spectrum (in dB)
figure;
plot(f/1e9, 20*log10(abs(X)))
xlim([2.3 2.8])    % Zoom in around expected sweep range
ylim([-100 100])
xlabel('Frequency (GHz)')
ylabel('Magnitude (dB)')
title('FFT of 1 ms Upconverted FMCW Sweep')
grid on

it then plotted this:

I am confused why is my RF FMCW starting at 2.5GHz while I clearly defined my fc = 2.45GHz?

Have you try using fs = 400 MHz instead of 200 MHz to see the impact on the output?

I don't have access to the phased array toolbox so I can't test your Matlab code as shown, but it seems to me you have some fundamental issues with sampling rates and center frequency. You generate the x vector of FMCW at a sample rate of 200MHz and then attempt to shift the frequency to 2.45GHz which is well beyond the sample rate and will cause all manner of aliasing issues that completely throw off the X axis on your FFT plot.

The issue is that the baseband sweep shape is not centered as you'd expected.

To get a sweep centered at 2.45 GHz, your baseband must sweep from -50 MHz → +50 MHz, which is exactly what SweepInterval = 'Symmetric' produces.